Aberrant DNA methylation in the promoter region of genes is found in a variety of human cancers and is thought to be associated with gene silencing. Restriction Landmark Genomic Scanning (RLGS) is currently the only technique that allows the scanning of thousands of promoter sequences for aberrant DNA methylation in human cancer. Aberrant DNA methylation was recently shown to be a major contributor and an early event in tumorigenesis, especially in the development of acute myeloid leukemia (AML). In this application we propose to investigate the role of DNA methylation in AML with special emphasis on clinical correlates. The samples that will be used for this study will come from the CALGB Leukemia Tissue Bank. Our hypothesis is that epigenetic changes (DNA methylation) are equally important as genetic alterations in leukemogenesis but have been underestimated in its extend. Since methylation changes could affect the transcription of genes it is likely that these epigenetic differences contribute to the molecular defects that underlie normal karyotype AML. Subsequently, aberrantly methylated targets can be used to identify novel diagnostic or prognostic biomarkers in AML. To test this hypothesis our specific aims are (1) to study methylation profiles in a subset of normal karyotype AML with blast counts >50 percent. (2) Rigorous statistical and bioinformatical analysis will identify diagnosis and relapse specific methylation events, candidate subclass predicting methylation events and finally methylation targets that correlate with clinical data such as duration of complete remission. (3) A small subset of highly informative methylation targets will be studied in detail by bisulfite sequencing. MS-PCR tests will be developed that allow (4) screening of larger patient samples. Statistical analysis will be performed to determine the value of a methylation event as a diagnostic biomarker or as a marker with predictive value.